Dissecting Reactivity of Clerocidin toward Common Buffer Systems by

The model drug clerocidin (CL) can form covalent adducts with both Tris and phosphate buffers with negative effects on biological activity, even thoug...
0 downloads 0 Views 153KB Size
Download PDF
Chem. Res. Toxicol. 2005, 18, 35-40

35

Articles Dissecting Reactivity of Clerocidin toward Common Buffer Systems by Means of Selected Drug Analogues Sara N. Richter,† Daniele Fabris,‡ Stefano Moro,† and Manlio Palumbo*,† Department of Pharmaceutical Sciences, University of Padova, via Marzolo 5, 35131 Padova, Italy, and Department of Chemistry and Biochemistry, University of Maryland, Baltimore County, Baltimore, Maryland 21250 Received May 21, 2004

The model drug clerocidin (CL) can form covalent adducts with both Tris and phosphate buffers with negative effects on biological activity, even though the latter is considered a largely inert physiological buffer. With the ultimate goal of learning how to control such reactivity and reduce undesired side reactions, we have investigated the influence of the different functionalities of CL on the formation of buffer adducts. For this reason, selected drug analogues were tested for their ability to react with the two buffers and comprehensive information was gained on both thermodynamics and kinetics aspects of these reactions. Two distinctive reactivity modes were readily observed. The first proved to be under kinetic control and involved the reaction of drug carbonyls (especially the aldehyde in C15) with the Tris amino group to form a Schiff base. The second was found to proceed under thermodynamic control through the attack at the oxirane ring of CL by the buffer’s nucleophilic groups (amino nitrogen in Tris and oxygen in phosphate). Important relationships between the two modes were noted, thus providing further demonstration that drug reactivity toward buffers cannot be directly predicted from the functionalities that are potentially involved in the initial reaction. On the contrary, as true for almost any structure bearing potentially reactive functionalities, any solid prediction should be based on a deeper understanding of the mutual influence of vicinal groups.

Introduction Promising drug candidates are subjected to intense investigation not only to assess their therapeutic potential but also to elucidate their mechanism of action. In general, in vitro studies precede those performed in vivo and provide the fundamental information necessary to ultimately approach clinical studies (1). Biological tests are usually performed using appropriate buffer systems that mimic physiological conditions (2-4). However, little attention is generally paid to possible side reactions between an active agent and a buffer, which may substantially affect the results of the assay by leading to possible drug deactivation. Furthermore, significant modifications of the buffer system could bear unexpected toxicological consequences by possibly decreasing the local buffering capacity or by generating novel drug conjugates of unknown biological activity (5-7). Similar considerations are also valid for in vivo systems that include true physiological buffers (e.g., phosphate, bicarbonate, creatine, etc.). We have recently completed a systematic analysis of the reactivity of the model drug clerocidin (CL)1 (8-14) toward specific buffers that are commonly used in drug testing and bioassays (15). This study has clearly proved that both Tris [tris(hydroxymethyl)aminomethane] and † ‡

University of Padova. University of Maryland.

phosphate buffers can provide stable covalent adducts with CL, even though the latter is generally considered inert when used in experiments requiring a physiological environment. This reactivity with buffer functionalities has benn shown to greatly affect the biological activity of CL, thus raising questions about its true therapeutic potential (15). CL is a natural product (16) that includes an R-dicarbonylic function in equilibrium with the hemiacetal form (Figure 1A), a strained epoxy ring in C12-C15, and an R,βunsatured aldehyde at position C3-C4 of the diterpenoid structure (16-18). While the reaction of the different functional groups with nucleophiles has been clearly described, there are several open questions on the possible interplay between vicinal groups, which may mutually affect their specific reactivities. The oxidation state of C14 and C15 may affect the strain introduced in the adjacent epoxide, thus modifying its reactivity toward nucleophiles. Conversely, the strained oxirane ring may influence the formation of hemiacetals or Schiff bases with the carbonyl functions present in the C12-C15 ring system. In addition, the diterpenoid portion may still induce an increase of the reactivity of CL, although the 1 Abbreviations: CL, clerocidin; NA, naphthalene; Tris, tris(hydroxymethyl)aminomethane; MS/MS, tandem mass spectrometry; TFA, trifluoroacetic acid; ESI, electrospray ionization; TOF, time-of-flight; FTMS, Fourier transform mass spectrometry; SAR, structure-activity relationship.

10.1021/tx049858i CCC: $30.25 © 2005 American Chemical Society Published on Web 12/09/2004

36

Chem. Res. Toxicol., Vol. 18, No. 1, 2005

Richter et al.

oxidation state of C14 in NA2 and NA5 and that of C14 and C15 in NA3 match those respectively found in spirocardin A and B, which are two closely related CL analogues with antibacterial activity (19). The synthetic derivatives were tested for their ability to react with Tris and phosphate buffers, and their adducts were unambiguously characterized by LC-MS and MS/MS analysis. A careful comparison of the reactivity exhibited by the CL analogues under identical experimental conditions allowed us to assess the mutually modulating effects exerted by the carbonyl functions of the C12-C15 ring and the epoxide moiety.

Experimental Procedures

Figure 1. Chemical structures of reacting compounds. (A) Lead compound, CL. The dicarbonylic form of the drug (DC) is shown in equilibrium with the closed hemiacetal species (HA). The numbering of atoms is given for the HA form of CL. The numbered atoms outline electrophilic carbons possibly involved in reactions with nucleophilic functions. (B) Intact epoxide NA derivatives. (C) Open epoxide NA derivatives. The carbonyl functions in NA1 and NA4 are unmodified; in NA2 and NA5, they are partially reduced; and in NA3, they are fully reduced. (D) Buffers under study. The molecules are shown in their neutral form. Molecular formulas, expected molecular weights, and pKa values for TRIS and phosphate are shown below each structure.

aldehyde in position 4 does not appear to be involved in adduct formation (14, 15). Synthetic CL derivatives were designed to address these questions, which required substitution of the diterpenoid moiety and modification of groups in the C12-C15 ring system (Figure 1) (14). Initially, an aromatic naphthalene (NA) was introduced in place of the diterpenoid portion to simplify the synthetic procedures and enable detection by fluorescence spectroscopy. Two new series of NA derivatives were then prepared with an intact (NA1, NA2, and NA3; Figure 1B) or open epoxide ring (NA4 and NA5; Figure 1C). Within each series, the oxidation state of the vicinal carbonyls was systematically changed to evaluate their effects on the epoxide reactivity. In particular, NA1 and NA4 included fully oxidized carbonyls similar to the original CL molecule. NA2 and NA5 contained partially reduced functions, which were fully reduced in NA3. It should be noted that the

Materials. CL was a kind gift of Leo Pharmaceutical Products (Ballerup, Denmark). NA derivatives were generously provided by Prof. S. Kobayashi (University of Tokyo, Japan). CL and NA analogues were dissolved in absolute ethanol, and concentrations were determined by measuring the absorbance at 230 and 282 nm, respectively, on a UV/vis Spectrometer Lambda 12 (Perkin-Elmer, MA), using the experimentally determined molar extinction coefficient of 11818 M-1 cm-1 for CL and 9493 M-1 cm-1 for NA derivatives. Working drug solutions were obtained by diluting fresh stocks in the appropriate buffer. Buffer components and chemicals for MS calibration were purchased from Sigma-Aldrich (MI). HPLC-grade solvents were obtained from LabScan (Dublin, Ireland). Preparation and Isolation of Drug-Buffer Adducts. CL and the NA series were dissolved in ethanol (50 µL) and diluted with 1 M Tris or phosphate buffers (pH 7.4) to give a final drug concentration of 1 mM. All reactions were carried out at 37 °C for different times, as indicated below. Reaction mixtures were filtered with 0.45 µm filters (Alltech, Ireland) and loaded on a HPLC C18 reverse phase column (Extent-C18, 4.6 mm × 250 mm, Agilent Technologies). The HPLC system consisted of the following: Series 200 Pumps (Perkin-Elmer, Italy); UV-1806, UV-Vis Detector (BIO RAD); NCI 900 Network Chromatography Interface (Perkin-Elmer). To separate the adducts, solvents A [CH3CN:H2O (9:1)/0.05% trifluoroacetic acid (TFA)] and B [H2O/0.05% TFA] were used. For CL and NA1 reactions, gradients were used as follows: A 20% for 8 min (equilibration time); A from 20 to 100% in 10 min; A from 100 to 20% in 2 min. For NA2: A 20% for 8 min (equilibration time); A from 20 to 60% in 10 min; A from 60 to 20% in 2 min. For NA3, NA4, and NA5: A 20% for 8 min (equilibration time); A from 20 to 40% in 10 min; A from 40 to 20% in 2 min. Peaks were detected at 230 nm, collected in Eppendorf tubes, dried at room temperature in Speed Vac UniVapo 100 H (UniEquip, Germany), and stored at 4 °C if necessary. Peak integrals were calculated with Turbochrom software (Perkin-Elmer). Aliquots from each mixture were analyzed by HPLC immediately after mixing and at regular intervals until reactions reached saturation or equilibrium. Adduct concentrations were calculated by peak integration, which gave the % of adducts formed in each reaction mixture. Kinetic constants for adduct formation were calculated according to the rate law of a pseudofirst order reaction, assuming that the concentration of buffer, which is present in large excess, would not change appreciably during reaction with each drug analogue. The applied rate equation was the following:

C ) CS ‚ (1 - e-k′‚t) with C and CS indicating the adduct concentrations at time t and saturation, respectively, and k′ representing the kinetic constant. MS and Tandem Mass Spectrometry (MS/MS) Analysis of NA-Buffer Covalent Adducts. LC-MS and direct infusion electrospray ionization (ESI) mass spectrometry were performed, respectively, on a time-of-flight (TOF) mass analyzer

Dissecting Clerocidin Reactivity toward Buffers Table 1. Kinetic Constants for CL and NA Derivatives with Tris and Phosphate Buffers k (min-1) sample CL NA1 NA2 NA3 NA4 NA5

Tris 0.22 ( 0.01 0.21 ( 0.01 7.6 ( 0.9 × 10-3 1.5 ( 0.6 × 10-4 0.047 ( 0.004 0.022 ( 0.002

Chem. Res. Toxicol., Vol. 18, No. 1, 2005 37 Table 2. m/z Values of Adducts Formed in the Reactions of CL and NA Derivatives with Tris and Phosphate Buffers m/z adduct

phosphate 1.8 ( 0.1 × 10-3 1.5 ( 0.4 × 10-3 0.011 ( 0.001 2.6 ( 0.1 × 10-4 N/A N/A

(Mariner ESI-TOF, Applied Biosystems, Foster City, CA) and on a Fourier transform mass spectrometer (FTMS), as described earlier (15). Computational Methods. Calculations were performed on a Silicon Graphics Octane R12000 workstation and on a 6 CPUs (PIV 2.0-3.0 GHZ) linux cluster running under openMosix architecture. Because the NA moiety is not directly involved in the chemical reactivity of the CL analogues, NA was replaced by a methyl group to expedite the calculations. All structures were constructed using the “Molecule Builder” module of Molecular Operating Environment (MOE 2002.03) (20) and were minimized using the MMFF94 force field (21, 22), until the rootmean-square deviation value provided by the Truncated Newton method was NA2 > NA3), but phosphate showed a different reactivity scale, with NA2 as the most reactive (NA2 > NA1 > NA3). Considering that Tris can either form reversible Schiff bases with carbonyls or irreversible epoxide adducts, while phosphate can only form the latter, the observed differences are not likely to be caused by the oxirane reactivity, which is common to both buffers, but rather by the different susceptibility of the carbonyls. This hypothesis was tested by evaluating the activation energy to reach the transition state and the stability of phosphate adducts with NA1 and NA2 (Figure 3). Quantum mechanical calculations performed on two simplified NA1- and NA2-phosphate adducts (see the Experimental Procedures) have shown that formation of the latter in aqueous solution is kinetically and thermodynamically more favorable than that of the former. The potential energies calculated for the NA2 transition state

Dissecting Clerocidin Reactivity toward Buffers

and final product in a reaction with H2PO4- were found to be ∼7 kcal/mol lower than those calculated for NA1. As clearly shown in Figure 3, both NA1- and NA2phosphate can provide intramolecular hydrogen bonding, but the latter forms a much stronger hydrogen bond between one of the phosphate oxygens and the hydroxyl group in C14 of the furan ring. This hydrogen bond stabilizes both the transition state (distance PO--HO = 2.1 Å) and the final product (distance PO--HO = 2.0 Å) and accounts for a more effective kinetic and thermodynamic stabilization of the NA2 adduct as compared to NA1phosphate.

Discussion The design of selected synthetic analogues has allowed us to dissect the reactivity of the natural product CL with common buffers, which has proven to substantially affect its biological activity (15). In particular, varying the oxidation state of the carbonyls present on the C12-C15 ring system has enabled the investigation of possible mutual effects between vicinal groups. The results obtained by reacting such analogues with Tris and phosphate buffers indicate the presence of two distinct types of reactivity. The first is characterized by a high reaction rate and involves the formation of reversible Schiff bases between the carbonyls and the amino group of Tris. The second is slower and irreversible and involves the nucleophilic attack of either the amino group of Tris nitrogen or the deprotonated oxygen of phosphate to the oxirane ring. The interplay between these two reaction modes, which determines the final outcome, should be carefully analyzed in terms of both kinetics and thermodynamics. Kinetic rates are determined first by the fast reversible reactions of the furanoid carbonyls and then by the slower reactivity of the oxirane group, which is in turn influenced by the carbonyls oxidation state. In fact, the spiro structure between the oxirane and the furanoid ring can only form if the aldehyde function in C15 can provide a hemiacetal to close the five-membered C12-C15 ring. Hence, the presence of the carbonyl at C15 decreases the activation energy of the attack to the epoxide group, thus increasing the corresponding reaction rate. Furthermore, the C14 carbonyl may turn the oxirane ring even more electrophilic through its electron withdrawing effects. From the thermodynamics point of view, while reversible reactions tend to reach equilibrium depending on the stability of the new adducts, irreversible reactions will eventually transform all of the initial drug into bufferdrug adducts. The reaction of NA1 with Tris may be favored by the contribution of the carbonyls, which provide reversible reactions with the primary amine and hydroxyl groups of Tris, combined with that of the spiro system, which increases the reactivity of the epoxide. Therefore, the overall stability of the final adduct can be explained by the irreversible nature of the product of epoxide reaction and by the relatively stable six-membered ring generated by the two reversible reactions involving the carbonyls. The adduct formed by NA2 lacks the extra stabilization introduced by the second reversible bond at C14, which leads to the six-membered ring. This feature also affects the reaction rate, which is slightly lower than that of NA1. NA3 lacks both contributions, as reflected by the much slower kinetics. However, the epoxide can still form the irreversible adduct and all of

Chem. Res. Toxicol., Vol. 18, No. 1, 2005 39

the initial NA3 is eventually converted into products at saturation. On the contrary, no epoxide function is present in NA4 and NA5; thus, only reversible adducts are observed; NA4 can still form a six-membered ring, which is sufficiently stable to be isolated and analyzed; NA5 provides only one reversible hemiacetal bond, which is remarkably labile and did not allow for full characterization. The oxirane ring has proven to be the only active group in the phosphate reaction. While possible attack at the carbonyls cannot be ruled out, such products are likely to be highly reversible and labile, as demonstrated by the fact that no stable adducts could be detected for the NA4 and NA5 analogues. Synthetic derivatives including an intact oxirane produced the expected adducts, but the observed reactivity scale (NA2 > NA1 > NA3) seemed to contradict at first sight the predictions based on the effects of vicinal groups on the oxirane. While both NA1 and NA2 enjoy the activating presence of the spiro system, the second vicinal carbonyl in NA1 should make the epoxide ring more electrophilic and increase the reactivity of this analogue. However, our quantum mechanical analysis show that the activation barrier is more favorable for NA2 than for NA1, thanks to the formation of a strong hydrogen bond between phosphate and the hydroxyl function of NA2 in the transition state (Figure 3). On one hand, this feature provides a plausible explanation for the unexpected reactivity scale; on the other, it also indicates that the NA2 adduct should be thermodynamically favored. The results provided by the selected analogues allow us to draw important conclusions on the reactivity of the natural product CL toward common buffers. The state of oxidation of the carbonyl functions at the C12-C15 ring and the presence of the epoxide group have important and interdependent effects on reactivity, while the contribution of the diterpenoid moiety is negligible. The consequences of opening the epoxide ring are limited to abolishing its own reactivity and do not interfere with adduct formation at the carbonyls. Vice versa, the carbonyls have proven to exert a major modulating influence on epoxide reactivity, which can be mediated by the formation of a spiro system, the electron withdrawing character of vicinal carbonyls, or both. The finely tuned modulation achieved in CL should provide an interesting case study for structure-activity relationship (SAR)driven drug design involving similar functional groups. In this direction, these results should also serve as a reminder that drug reactivity cannot be completely predicted from the identity of the primary functional groups involved in a certain reaction but requires a deeper understanding of the influence of neighboring groups. At the functional level, the consequences for CL and similar natural/synthetic products that may be involved in unwanted buffer interactions extend beyond the possible misevaluation of their biological activity in vitro but may potentially include unpredictable effects on their therapeutic activity and toxicity in vivo. The investigation of the mechanism of action of such drugs should not be limited to their possible interactions with the customary group of nucleophilic species present in cellular/extracellular environments, including nucleic acids (13, 14), proteins, etc. The possible formation of adducts with physiological buffers, which may present unknown activity and metabolism or may decrease the local buffering

40

Chem. Res. Toxicol., Vol. 18, No. 1, 2005

capacity, should be added to the list of potentially negative effects to be carefully investigated.

Acknowledgment. This work has been carried out with the financial support of Associazione Italiana per la Ricerca sul Cancro (AIRC), Milan, and the Italian Ministry for University and Research (MIUR), Rome, Italy. Supporting Information Available: MS/MS analysis of NA1-Tris bis and monoadducts and NA4-Tris, NA5-Tris, and NA3-Tris monoadducts. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Pritchard, J. F., Jurima-Romet, M., Reimer, M. L., Mortimer, E., Rolfe, B., and Cayen, M. N. (2003) Making better drugs: Decision gates in nonclinical drug development. Nat. Rev. Drug Discovery 2, 542-553. (2) Ferguson, W. J., Braunschweiger, K. I., Braunschweiger, W. R., Smith, J. R., McCormick, J. J., Wasmann, C. C., Jarvis, N. P., Bell, D. H., and Good, N. E. (1980) Hydrogen ion buffers for biological research. Anal. Biochem. 104, 300-310. (3) Good, N. E., and Izawa, S. (1972) Hydrogen ion buffers. Methods Enzymol. 24, 53-68. (4) Good, N. E., Winget, G. D., Winter, W., Connolly, T. N., Izawa, S., and Singh, R. M. (1966) Hydrogen ion buffers for biological research. Biochemistry 5, 467-477. (5) Nassar, A. E., and Talaat, R. E. (2004) Strategies for dealing with metabolite elucidation in drug discovery and development. Drug Discovery Today 9, 317-327. (6) Nassar, A. E., and Lopez-Anaya, A. (2004) Strategies for dealing with reactive intermediates in drug discovery and development. Curr. Opin. Drug Discovery Dev. 7, 126-136. (7) Bugrim, A., Nikolskaya, T., and Nikolsky, Y. (2004) Early prediction of drug metabolism and toxicity: Systems biology approach and modeling. Drug Discovery Today 9, 127-135. (8) Kawada, S., Yamashita, Y., Fujii, N., and Nakano, H. (1991) Induction of a heat-stabletopoisomerase II-DNA cleavable complex by nonintercalative terpenoides, terpentecin and clerocidin. Cancer Res. 51, 2922-2925. (9) Kawada, S. Z., Yamashita, Y., Uosaki, Y., Gomi, K., Iwasaki, T., Takiguchi, T., and Nakano, H. (1992) UCT4B, a new antitumor antibiotic with topoisomerase II mediated DNA cleavage activity, from Streptomyces sp. J. Antibiot. Tokyo 45, 1182-1184.

Richter et al. (10) Kawada, S., Yamashita, Y., Ochiai, K., Ando, K., Iwasaki, T., Takiguchi, T., and Nakano, H. (1995) Terpentecin and UCT4B, new family of topoisomerase II targeting antitumor antibiotics produced by Streptomyces: Producing organism, fermentation and large scale purification. J. Antibiot. Tokyo 48, 211-216. (11) Binaschi, M., Zagotto, G., Palumbo, M., Zunino, F., Farinosi, R., and Capranico, G. (1997) Irreversible and reversible topoisomerase II DNA cleavage stimulated by clerocidin: Sequence specificity and structural drug determinants. Cancer Res. 57, 1710-1716. (12) Borgnetto, M. E., Tinelli, S., Carminati, L., and Capranico, G. (1999) Genomic sites of topoisomerase II activity determined by comparing DNA breakage enhanced by three distinct poisons. J. Mol. Biol. 285, 545-554. (13) Gatto, B., Richter, S., Moro, S., Capranico, G., and Palumbo, M. (2001) The topoisomerase II poison clerocidin alkylates nonpaired guanines of DNA: implications for irreversible stimulation of DNA cleavage. Nucleic Acids Res. 29, 4224-4230. (14) Richter, S., Gatto, B., Fabris, D., Takao, K., Kobayashi, S., and Palumbo, M. (2003) Clerocidin alkylates DNA through its epoxide function: Evidence for a fine-tuned mechanism of action. Nucleic Acids Res. 31, 5149-5156. (15) Richter, S., Fabris, D., Binaschi, M., Gatto, B., Capranico, G., and Palumbo, M. (2004) Effects of common buffer systems on drug activity: The case of clerocidin. Chem. Res. Toxicol. 17, 492-501. (16) Andersen, N. R., Lorck, H. O., and Rasmussen, P. R. (1983) Fermentation, isolation and characterization of antibiotic PR1350. J. Antibiot. Tokyo 36, 753-760. (17) Andersen, N., Lork, H., and Rassmussen, P. (1984) The relative and absolute configuration of clerocidin and its cometabolites. Tetrahedron Lett. 25, 469-472. (18) Andersen, N., and Rasmussen, P. (1984) The costitution of clerocidin. A new antibiotic isolated from Oidiodendron truncatum. Tetrahedron Lett. 25, 465-468. (19) Nakajima, N., Okazaki, T., Iwado, S., Kinoshita, T., and Haneishi, T. (1989) New diterpenoid antibiotics, Spirocardins A and B. J. Antibiot. Tokyo 42, 1741-1748. (20) Molecular Operating Environment (MOE 2002.03), Chemical Computing Group, Inc., Montreal, Quebec, Canada. (21) Halgren, T. (1996) Merck molecular force field. IV. Conformational energies and geometries for MMFF94. J. Comput. Chem. 17, 587615. (22) Halgren, T. (1999) MMFF VII. Characterization of MMFF94, MMFF94s, and other widely available force fields for conformational energies and for intermolecular-interaction energies and geometries. J. Comput. Chem. 20, 730-748. (23) Spartan O2, Wavefuntion Inc., Irvine, CA.

TX049858I